Good Strategy Bad Strategy

Richard Rumelt’s central argument is that strategy is not a statement of ambition. It is a response to a consequential challenge. A growth target, an inspiring vision, a list of values, and a crowded roadmap may all be useful, but none says what obstacle matters most or how an organization intends to overcome it. Calling them strategy hides the choice that strategy exists to make.

Good strategy has a simpler structure. It interprets the situation, chooses a general way through it, and coordinates action around that choice. Rumelt calls these three elements the kernel: a diagnosis, a guiding policy, and coherent actions. The kernel matters because each part corrects a different planning failure. Diagnosis prevents a team from acting on a confused picture of reality. Guiding policy prevents analysis from ending without a choice. Coherent action prevents the choice from dissolving into disconnected activity.

A diagnosis reduces complexity without pretending it is simple

A diagnosis explains why the situation is difficult. It selects the features that matter and organizes them into a tractable account of the challenge. This is not merely a description. “Revenue is below plan” names an outcome; it does not explain what is producing it. A diagnosis might instead identify that a product solves a frequent problem but requires users to change a trusted workflow before they can experience its value. That interpretation directs attention toward switching risk rather than toward every possible way to increase revenue.

Any diagnosis is an act of judgment. Complex situations support many true descriptions, but strategy needs the one that exposes leverage. Rumelt compares this work to medical diagnosis: the physician does not respond to every symptom independently. A useful diagnosis connects symptoms to a cause or condition and thereby rules some treatments in and others out. The analogy does not imply certainty. A diagnosis is a working explanation that action can test.

This is why disagreement at the diagnostic level is valuable. Two people can endorse the same goal while holding incompatible theories of the obstacle. One believes acquisition is weak; another believes new users arrive but fail to recognize value. Their proposed initiatives may look like a priority conflict, but the deeper conflict concerns what is true. Making each diagnosis explicit allows evidence to resolve the disagreement. Combining both sets of initiatives usually avoids the argument while also dispersing effort.

A guiding policy makes a choice without prescribing every move

Once the challenge is named, the guiding policy establishes the overall approach. It is more specific than a slogan and less specific than a project plan. “Delight customers” offers no guidance because almost any action can be justified by it. “Reduce the need for customers to trust us before they see value” constrains the field: it favors reversible trials, visible proof, and low-risk adoption over features that deepen the commitment required at the start.

The policy is “guiding” because people still need judgment. A sound strategy does not anticipate every local decision. It creates a common criterion that lets many decisions point in the same direction. It also creates boundaries. A policy that cannot reject an attractive project is probably too vague to coordinate action.

Rumelt’s account of Apple in 1997 shows this kind of choice. When Steve Jobs returned, Apple had a sprawling product line, licensing arrangements, peripheral businesses, and severe financial pressure. Jobs cut models and activities, concentrated the company on a small product set, and stabilized its core. Rumelt emphasizes that the striking move was not a grand forecast of the future. It was recognizing the company’s immediate strategic problem and focusing resources so Apple could survive long enough to exploit later opportunities. The example demonstrates that strategy can begin with concentration and repair rather than novelty.

Coherence turns ordinary actions into strategic force

The third element is coordinated action. Individual actions do not become strategic because they appear in a strategy document. They become strategic when they implement the guiding policy and strengthen one another. A lower-friction trial, proof from a credible peer, and an onboarding path that produces value before migration could form a coherent response to switching risk. Adding an unrelated brand refresh and a new enterprise tier might each be defensible, but they do not gain force from the same diagnosis.

Coherence is a source of advantage because coordination is difficult to copy. A competitor can imitate one feature. It is harder to reproduce a whole system in which product design, distribution, pricing, capabilities, and operating choices fit together. Rumelt’s idea is not that every company needs an elaborate interlocking system. It is that a collection of reasonable initiatives is weaker than a smaller set that creates cumulative pressure on the chosen obstacle.

This also explains why prioritization alone is not strategy. Ranking twenty requests produces an order of execution, but not necessarily a theory of progress. The stronger test is relational: if this action succeeds, does it make the other chosen actions more effective? If removing an initiative leaves the strategic logic unchanged, it may be useful work, but it is not part of the strategy’s core.

Bad strategy protects people from making hard choices

Rumelt identifies several recurring forms of bad strategy. Fluff uses inflated language or fashionable abstractions in place of insight. Failure to face the challenge declares a desired future without naming what blocks it. Mistaking goals for strategy treats an outcome such as growth or leadership as though wanting it explains how to obtain it. Bad strategic objectives produce a long, impractical list or ask for progress without addressing the diagnosed obstacle.

These are not merely writing defects. They often perform a social function. A long list allows every constituency to keep its priority. A lofty goal preserves agreement because it postpones conflict over means. Ambiguous language lets people interpret the strategy in mutually incompatible ways. Bad strategy can therefore feel more comfortable than good strategy. It offers aspiration without exposing the tradeoffs, uncertainty, and power involved in concentration.

The language test is useful but not sufficient. Replacing “world-class innovation” with plainer words does not create a strategy. The substantive test is whether the plan contains a contestable explanation of the challenge and a choice that redirects resources. If a competitor could adopt the statement without changing anything, or if every current initiative remains equally important after the strategy is announced, the hard part is still missing.

Leverage comes from concentrating effort at the right point

Good strategy looks for leverage: a point where focused effort produces an outsized change. Leverage may come from anticipating a shift, exploiting an asymmetry, combining capabilities, or resolving a bottleneck that constrains the whole system. The strategic insight lies in seeing why this point matters now.

Concentration carries an important qualification. Focus is not automatically wise. Concentrating on the wrong diagnosis makes error more expensive. The remedy is not to keep every option equally funded; that eliminates strategic force. It is to state the diagnosis clearly enough that evidence can challenge it, choose actions that reveal whether it is right, and update when reality contradicts the theory.

Rumelt uses the idea of a proximate objective for situations where the ultimate goal is too distant or uncertain to guide present action. A proximate objective is close enough to be feasible with current knowledge while valuable enough to improve the strategic position. It turns ambiguity into a solvable next challenge. A team may not know how to dominate a category, but it may be able to become indispensable to one narrow group whose problem it understands unusually well.

The proximate objective is not a timid substitute for ambition. It is a way to sequence it. Achievement changes what is knowable and possible. The next diagnosis can be made from a stronger position, with evidence that was unavailable before action.

Strategy is a hypothesis that earns commitment

The kernel makes strategy sound orderly, but Rumelt does not present it as a mechanical planning recipe. Strategic work involves insight, analogy, judgment, and the design of a response under uncertainty. The diagnosis and guiding policy are hypotheses about how the situation works. Coherent actions commit enough effort to test those hypotheses in the world.

That combination—commitment with revisability—is the book’s most useful discipline. Refusing to choose until uncertainty disappears produces drift. Treating a strategy as permanent truth produces rigidity. A team should be able to say: this is the challenge we currently believe matters most; this is the approach we have chosen because of that belief; these actions express the approach and reinforce one another; these results would make us revise the diagnosis.

The practical standard is therefore stricter than having a plan. A strategy is good when it creates a clear explanation of the obstacle, an exclusionary approach, and concentrated action whose results teach the organization whether its explanation was right. It replaces the comfort of many ambitions with a coherent bet about what matters.